Automated guided vehicle path planning by dynamically adjusting ant colony algorithm heuristic factor

Abstract

In view of the problems of traditional ant colony (ACO) algorithm, such as slow convergence speed, many iterations and easy to fall into local optimal, a dynamically ant colony optimization(DACO) algorithm was proposed. With the optimal path as reference, experiments were carried out on the values of pheromone heuristic factor and expectation heuristic factor of the traditional ant colony algorithm. When the value range of α is [1,3] and the value range of β is [7,9], the shortest path can be obtained. In view of the above range of values, the experiments on how to value the two parameters show that when α follows the normal function distribution curve and β follows the symmetric curve of α with respect to y=10, the convergence speed of the algorithm is accelerated and the number of iterations is reduced, so as to avoid the ant falling into the local optimum and thus failing to find the optimal solution. ROS robot was used to verify the experimental platform, and the results show that the optimization time of the improved algorithm is 6.05% shorter than that of the traditional ant colony algorithm.

Keywords

• ant colony algorithm
• automated guided vehicle(agv) path
• pheromone heuristic factor
• expectation heuristic factor
• optimal path